10 apps in 20 days, with Claude Code

I've been a Mac user and a UI obsessive for as long as I can remember, but I've never built an app. No programming background or head for code beyond HTML and CSS. I'd been watching AI since GPT-3 with the vague, distant amazement of someone who grasps how impressive something is but without an understanding of its importance. Less than a month ago, that started changing. I sat down with Claude Code and started building a Mac app, an idea for a visual CSS editor I had for a decade, just to see what would happen. A month later, to my astonishment, I had completed ten apps, some small, some bigger. I'm still kind of stunned.

It was the most engrossed I've been in anything in years. I had to force myself to close my MacBook at night and could barely wait to open it again in the morning.

What surprised me most was how truly collaborative the process felt. I was the visionary, art director, and the QA tester. Claude was the implementor and the sounding board. Good results still came down to my taste, my attention to detail, my ability to spot the thing that was almost-but-not-quite right. But Claude got us 70% of the way there on the first or second ask, almost every time. Claude's ability to say "oh yeah, I get what he's going for" is what made the pace possible. I never once opened Xcode, never once sketched out a UI.

These are good apps, I think. Considered. Polished. The kind of small, opinionated tools I've always wished existed exactly the way I wanted them to. But here's the part I keep turning over: none of them are changing anyone's life. They fill my particular desires for how a teleprompter or a bookmark manager or a CSS editor should work. They're not solving unmet needs at any meaningful scale. So on one hand, this is an incredible personal milestone, and it does feel like a small confirmation of the societal shift everyone keeps talking about. On the other hand, I have to ask honestly: is what I made, however much joy it brought me, just another flavor of AI slop? Am I just doing this because I am now able to and had a lot of free time recently? So I'm kind of back where I started: once again amazed by the possibilities and once again unsure about the impact.

Though, ironically, given my design focus, what I do think may have the most longterm practical impact is the process automation AI brings. The way Claude zipped through the notarization process, fixing problems it encountered while uploading and organizing files, instantly updating app documentation as we released updates, was almost as impressive as the app creation process itself. That's the kind of thing that feels meaningful at scale, versus the more flashy "everyone's a creator / entrepreneur now" headlines. The app store process itself would have stopped me cold in my tracks had I been going it alone. Maybe that says more about me than Claude, though.